济南市南部山区崮山流域地处我国北方土石山地丘陵区，以水力侵蚀为主的土壤侵蚀现象严重且生态环境脆弱，径流预测是水土保持监测和预报的重要基础，精度较高的的径流计算模型可以为济南市南部山区更好地开展水土保持工作提供技术支撑与数据基础。根据径流曲线模型（SCS-CN）原理和崮山流域内5个雨量站、1个水文站近10年的实测降雨、径流资料,借助ArcGIS平台利用优度拟合统计分析法及Nash-Sutcliffe效率系数验证法对模型参数初损率（λ）和径流曲线数（CN）进行了优化检验，结果表明：参数优化后的模型精确度较高（实测值与计算值分析结果为回归直线斜率K=0.905 8、确定系数R2=0.812 7、纳什效率系数ENS=0.796 9）可以更好地适用于崮山流域径流测算；对2019年崮山流域29次侵蚀性降雨进行降雨产流估算，并累加计算得流域年径流量为0.53亿 m3，年径流深处于34.15~371.52 mm，年均径流深为134.52 mm，汛期降雨产生的径流量占年径流量的90.27%。

Gushan watershed in the southern mountainous area of Jinan is located in the northern China.Soil erosion is a serious problem,dominated by hydraulic erosion and the ecological environment is fragile.Runoff carrying sediment are intuitive manifestations of soil erosion.The Gushan watershed is a key area for dynamic monitoring of water and soil erosion at the provincial level in Jinan.Complete rainfall and runoff datasets and the parameter optimization of the rainfall-runoff estimation model (runoff curve model (SCS-CN)) can improve the applicability of the model in the watershed and provide technical support and basis for the monitoring and forecasting of soil and water conservation in the region. Based on multi-source remote sensing technology and ArcGIS 10.7 platform,and according to the principle of the SCS-CN model,comprehensive consideration is given to the soil types in the watershed,the degree of early soil moisture,land use types,and topographic features.Measured rainfall and runoff data of five rainfall stations and one hydrological station in the basin for the past 10 years is selected,and goodness-fitting statistical analysis method and Nash-Sutcliffe efficiency coefficient verification method are used to determine the initial loss rate (λ) and the number of runoff curves of the model parameters (CN) to carry out optimization test.The rainfall-runoff estimation results of the standard SCS-CN model,Woodward model,and optimized SCS-CN model in the watershed are used to improve the estimation accuracy of rainfall-runoff of the SCS-CN model in the Gushan watershed. (1) When the value of the model parameter λ is less than 0.06,the runoff simulation result fits well with the measured value.After parameter optimization,λ is selected as 0.05.(2) The runoff curve model is greatly affected by topographical changes in the limestone mountain area.After using CN value slope conversion model,the accuracy of the model is further improved.(3) The optimized model has high accuracy with K=0.905 8,R2=0.812 7,ENS=0.796 9 which can be better applied to the rainfall-runoff estimation in the Gushan watershed.(4) The rainfall-runoff generation of 29 erosive rainfalls in 2019 is estimated in the Gushan watershed,and the cumulative calculations show that the annual runoff in the watershed is 0.53×10.8 m3,the annual runoff depth is between 34.15 mm and 371.52 mm,and the average annual runoff depth is 134.52 mm,respectively.The runoff generated by rainfall during the flood season accounts for 90.27% of the annual runoff. The optimized SCS-CN model has higher applicability in the study area.The model goodness of fit determination coefficient is increased by 17.68% compared with the Woodward model,and is increased by 113.08% compared with the standard SCS-CN model.The Nash efficiency coefficient (ENS) is compared with the Woodward model shows an increase of 48.81%,and 137.59% compared to the standard SCS-CN model.(2) The runoff generated by the various erosive rainfall events in the study area is affected by many factors such as the amount of rainfall and vegetation coverage.Based on research results,it can be determined that rainfall and runoff are positively correlated,and the degree of vegetation coverage and runoff has a negative correlation.(3) The annual runoff of the Gushan watershed was 0.53×10.8 m3 in 2019,and the spatial distribution showed a pattern of high in the north and low in the south,high in the east,and low in the west.The order of the average annual runoff depth of each administrative township in the basin is Guyunhu Sub-district>Zhangxia Sub-district>Wande Sub-district.(4) The optimization of the parameter CN value in the model optimization process is obtained based on the look-up table value.The soil type,the artificially cultivated tree species,natural tree species,and other vegetation types are staggered and complicated.The next research plan is to establish a runoff community and to further optimize the parameter values of the SCS-CN model and improve its application accuracy in the southern mountainous area of Jinan.